Power transformation equipment fault risk assessment system based on intelligent perception

Abstract

The invention provides a power transformation equipment fault risk assessment system based on intelligent perception, which comprises a data acquisition module, an intelligent diagnosis module, a fault early warning module and a decision support module, and is characterized in that the data acquisition module is connected with the intelligent diagnosis module, and the intelligent diagnosis module is connected with the fault early warning module and the decision support module. According to the invention, an intelligent diagnosis model is constructed by adopting a multi-classifier ensemble learning method, so that different types of faults can be accurately identified, and the diagnosis precision is improved by analyzing historical fault data; based on a fault prediction result, specific maintenance suggestions and priority ranking of maintenance work are provided for users, and stable operation of a power system is guaranteed. The system not only improves the fault processing efficiency, but also enhances the safety and reliability of the power system.

Description

Technical Field

[0001] The present invention belongs to the technical field of substation equipment failure maintenance management, and specifically relates to a substation equipment failure risk assessment system based on intelligent perception. Background Art

[0002] With the continuous expansion of the scale of power systems and the growing demand for electricity, the stable operation of substation equipment is crucial to the reliability and safety of the entire power system. Failure of substation equipment may lead to power outages and significant economic losses. Therefore, the need for comprehensive perception of equipment status is becoming increasingly urgent. Traditional sensing technology and fault diagnosis technology can no longer meet the application needs of modern power grids, and the development of intelligent sensing technology provides a new solution for equipment status monitoring.

[0003] Traditional fault diagnosis methods are mainly based on empirical rules and simple data analysis, which have certain bottlenecks, specifically reflected in:

[0004] 1. Low data utilization rate. A large amount of data is generated during the operation of substation equipment, but this data is often not effectively utilized. In addition, the current data analysis method is single and cannot fully reflect the operating status and health status of the equipment.

[0005] 2. Lack of predictive maintenance. Traditional maintenance methods are mainly passive, that is, repairs are performed after a fault occurs, and preventive maintenance cannot be achieved. The existing fault warning system usually makes predictions based on historical operating fault data training models, but lacks the ability to predict special situations. Summary of the invention

[0006] The purpose of the present invention is to provide a substation equipment failure risk assessment system based on intelligent perception to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solutions: a substation equipment fault risk assessment system based on intelligent perception, comprising a data acquisition module, an intelligent diagnosis module, a fault warning module, and a decision support module.

[0008] Data acquisition module: collects the operating status data of the substation equipment in real time based on the real-time data of the substation equipment and sensor data;

[0009] Intelligent diagnosis module: Based on the intelligent diagnosis analysis model of multi-classifier ensemble learning, it can identify the fault type and perform fault diagnosis based on historical fault information;

[0010] Fault warning module: When potential risks are detected, early warning information is issued in a timely manner to notify maintenance personnel to take action;

[0011] Decision support module: Provides specific maintenance suggestions and priority rankings based on fault prediction results, helping managers develop more effective maintenance plans and optimize resource allocation;

[0012] The data acquisition module is connected to the intelligent diagnosis module, and the intelligent diagnosis module is respectively connected to the fault warning module and the decision support module.

[0013] Furthermore, the data acquisition module is responsible for collecting and transmitting the status information of the equipment in real time, including equipment identification data, communication status data, electrical parameter data, physical parameters, environmental conditions, operating parameters and maintenance records.

[0014] Furthermore, the intelligent diagnosis module is divided into a base model and a meta-classifier according to the data type, wherein the base model selects different basic models for training and prediction according to the data set, and the meta-model is modeled and trained according to the prediction results of the base model, and the relationship between the weights of the learning prediction results and the model results is learned to maximize the accuracy of the overall model. The weights of the learning prediction results are implemented by logistic regression and weighted average, and the relationship between the learning model results is implemented by tree model and neural network.

[0015] Furthermore, the base model is verified according to different data subsets, 80% of the data subsets are divided according to the data type and data relevance, and 20% of the data subsets are randomly divided; for data subsets of stream data type, long short-term memory network is used as the base model, and for other types of data subsets, multilayer perceptron is used as the base model.

[0016] Furthermore, the prediction results of the base models are combined into a new feature matrix. For M base models h1,h2...h m ,According to the prediction results of each base model, a new feature matrix is ​​constructed as the training data of the meta-model.

[0017] Furthermore, the fault warning module needs to model the fault information, including circuit breaker fault, disconnector fault, transformer fault, arrester fault, capacitor bank fault, control and protection equipment fault, fault caused by environmental factors, fault caused by improper operation and maintenance, line fault, communication fault, and generate a labeled training data set D = {(x1, y1), (x2, y2)...., (x n ,y n )}, where x i is the prediction result of the base model, y i is the corresponding fault label.

[0018] Furthermore, the fault warning module first uses the base model to predict the data according to the transmission data of the data acquisition module. After obtaining the prediction result of each base model, it uses the meta-model to calculate its similarity with each sample in the training set, and finds the k most similar training samples. According to the labels of these k neighbors, the distance weighted voting method is used to predict the fault information of the new sample.

[0019] Furthermore, the similarity is calculated as follows:

[0020]

[0021] Where d is the prediction result of similarity, m is the prediction result of the base model, is the prediction result of the base model of the collected data, δ is the indicator function, x i To collect data.

[0022] Furthermore, the distance weighted voting method is calculated as follows:

[0023]

[0024] Where N k (x) represents the set of k nearest training samples, w i is the weight of the ith neighbor, c is the corresponding fault type, Indicates the predicted category, y i Indicates the category to which the i-th sample in the training set belongs, w i The calculation method is as follows:

[0025]

[0026] Furthermore, the decision support module establishes an operation and inspection strategy library, forms a one-to-one special strategy or a one-to-many progressive strategy according to the complexity of equipment defects and failures, and provides specific maintenance suggestions and priority ranking according to the failure prediction results.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. Real-time perception and condition maintenance of equipment: The system supports the transformation of substation equipment from periodic maintenance to condition maintenance, provides basic data support for the comprehensive optimization and management of the equipment life cycle based on condition monitoring, reduces the probability and time of equipment power outage maintenance, and further reduces operating costs.

[0029] 2. Multi-dimensional state perception. The present invention performs intelligent perception of equipment based on an intelligent diagnosis and analysis model based on multi-classifier ensemble learning, wherein the base model is modeled not only for the correlation of data, but also randomly selects data from the data set for modeling to improve the robustness of the model.

[0030] 3. Improve the accuracy and timeliness of fault detection. By installing various sensors and intelligent sensing devices, the system can collect multi-dimensional data such as temperature, vibration, current, voltage, etc. of the equipment in real time to ensure the integrity and accuracy of the data.

[0031] 4. Improve the scientificity and effectiveness of decision support, automatically generate optimization suggestions and action guidelines based on data analysis results, and help decision makers formulate specific action plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0033] Figure 1 The present invention is a schematic diagram of the structure of a power transformation equipment failure risk assessment system based on intelligent perception.

[0034] Figure 2 The present invention is a method flow chart of a system for assessing the risk of power transformation equipment failure based on intelligent perception. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Reference Figure 1 As shown, a substation equipment failure risk assessment system based on intelligent perception includes: a data acquisition module, an intelligent diagnosis module, a fault warning module, and a decision support module.

[0037] The data acquisition module is connected to the intelligent diagnosis module, and the intelligent diagnosis module is connected to the fault warning module and the decision support module.

[0038] Data acquisition module: mainly collects the operating status data of substation equipment in real time based on the real-time data of substation equipment and sensor data;

[0039] The data acquisition module is connected to the intelligent diagnosis module, and the intelligent diagnosis module is respectively connected to the fault warning module and the decision support module.

[0040] In a specific embodiment of the present invention, the data acquisition module is responsible for collecting and transmitting the status information of the device in real time, including device identification data, communication status data, electrical parameter data, physical parameters, environmental conditions, operating parameters and maintenance records.

[0041] It should be noted that equipment identification data includes equipment serial number, model, and manufacturer information; communication status data includes communication interface status and network connection status; electrical parameter data includes voltage, current, power, and frequency of substation equipment; physical parameters include equipment temperature, vibration, noise, and humidity; environmental conditions include equipment location, ambient temperature and humidity, and atmospheric pressure; operating parameters and maintenance records include equipment switch status, operating time, and start and stop times.

[0042] The present invention integrates multi-source data by collecting, integrating, cleaning, converting and balancing the equipment identification data, communication status data, electrical parameter data, physical parameters, environmental conditions, operating parameters and maintenance records of the substation equipment in the data acquisition module, thus providing a good data foundation for subsequent models.

[0043] Intelligent diagnosis module: Based on the intelligent diagnosis analysis model of multi-classifier ensemble learning, it identifies the fault type and performs fault diagnosis based on historical fault information.

[0044] In a specific embodiment of the present invention, the intelligent diagnosis module is divided into a base model and a meta-classifier according to the data type, wherein the base model selects different basic models for training and prediction according to the data set, and the meta-model is modeled and trained according to the prediction results of the base model, and the relationship between the weights of the learning prediction results and the model results is learned to maximize the accuracy of the overall model. The weights of the learning prediction results are implemented by logistic regression and weighted average, and the relationship between the learning model results is implemented by tree model and neural network.

[0045] It should be noted that the base model is verified according to different data subsets, 80% of the data subsets are divided according to the data type and data relevance, and 20% of the data subsets are divided randomly; for the data subsets of the stream data type, the long short-term memory network is used as the base model, and for other types of data subsets, the multilayer perceptron is used as the base model.

[0046] It should be noted that the prediction results of the base models will be combined into a new feature matrix. For M base models h1,h2...h m ,According to the prediction results of each base model, a new feature matrix is ​​constructed as the training data of the meta-model.

[0047] The present invention utilizes advanced data analysis and artificial intelligence technologies in the intelligent diagnosis module to automatically detect and diagnose abnormal conditions of substation equipment, which can detect potential faults or abnormalities early and prevent small problems from turning into major faults. At the same time, by analyzing large amounts of data, the type and location of the fault can be accurately identified, helping technicians to quickly locate the problem.

[0048] Fault warning module: When a potential fault risk is detected, a warning message is issued in time to notify maintenance personnel to take action.

[0049] In a specific embodiment of the present invention, the fault warning module needs to model fault information, including circuit breaker failure, disconnector failure, transformer failure, lightning arrester failure, capacitor bank failure, control and protection equipment failure, failure caused by environmental factors, failure caused by improper operation and maintenance, line failure, and communication failure.

[0050] It should be noted that, according to the prediction results of the fault occurrence time-based model, a labeled training data set D = {(x1, y1), (x2, y2)...., (x n ,y n )}, where x i is the prediction result of the base model, y i is the corresponding fault label.

[0051] It should be noted that the fault warning module first uses the base model to predict the data based on the transmission data of the data acquisition module. After obtaining the prediction results of each base model, it uses the meta-model to calculate its similarity with each sample in the training set, and finds the k most similar training samples. According to the labels of these k neighbors, the distance weighted voting method is used to predict the fault information of the new sample.

[0052] It should be noted that the calculation method of using the meta-model to calculate the similarity between the collected data and the samples in the training set is as follows:

[0053]

[0054] Where d is the prediction result of similarity, m is the prediction result of the base model, is the prediction result of the base model of the collected data, δ is the indicator function, x i To collect data.

[0055] It should be noted that the distance weighted voting method is calculated as follows:

[0056]

[0057] Where N k (x) represents the set of k nearest training samples, w iis the weight of the ith neighbor, c is the corresponding fault type, Indicates the predicted category, y i Indicates the category to which the i-th sample in the training set belongs, w i The calculation method is as follows:

[0058]

[0059] The present invention can predict the operating status of the equipment by real-time monitoring and analyzing the operating data of the equipment in the fault warning module, thereby providing important support in early warning, fault location, preventive maintenance and performance monitoring, improving equipment reliability and reducing equipment downtime.

[0060] Decision support module: Provides specific maintenance suggestions and priority rankings based on fault prediction results, helping managers develop more effective maintenance plans and optimize resource allocation.

[0061] In a specific embodiment of the present invention, the decision support module establishes an operation and maintenance strategy library, forms a one-to-one special strategy or a one-to-many progressive strategy according to the complexity of equipment defects and failures, and provides specific maintenance suggestions and priority ranking according to the fault prediction results.

[0062] The present invention uses multi-dimensional data analysis in the decision support module to customize and adjust maintenance suggestions according to different fault scenarios with data-driven, helping decision makers evaluate different decision plans. It can not only process large amounts of complex data, make predictions and analyses, and improve decision efficiency and quality, but also perform real-time analysis and predictions to provide real-time decision support.

[0063] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A substation equipment failure risk assessment system based on intelligent perception, characterized in that: include: Data acquisition module: collects the operating status data of the substation equipment in real time based on the real-time data of the substation equipment and sensor data; Intelligent diagnosis module: Based on the intelligent diagnosis analysis model of multi-classifier ensemble learning, it can identify the fault type and perform fault diagnosis based on historical fault information; Fault warning module: When potential risks are detected, early warning information is issued in a timely manner to notify maintenance personnel to take action; Decision support module: Provides specific maintenance suggestions and priority rankings based on fault prediction results, helping managers develop more effective maintenance plans and optimize resource allocation The data acquisition module is connected to the intelligent diagnosis module, and the intelligent diagnosis module is respectively connected to the fault warning module and the decision support module.

2. According to the intelligent perception-based substation equipment failure risk assessment system of claim 1, it is characterized in that: The data acquisition module is responsible for collecting and transmitting the status information of the equipment in real time, including equipment identification data, communication status data, electrical parameter data, physical parameters, environmental conditions, operating parameters and maintenance records.

3. According to the intelligent perception-based substation equipment failure risk assessment system of claim 1, it is characterized in that: The intelligent diagnosis module is divided into a base model and a meta-classifier according to the data type, wherein the base model selects different basic models for training and prediction according to the data set, and the meta-model is modeled and trained according to the prediction results of the base model, and the relationship between the weights of the learning prediction results and the model results is learned to maximize the accuracy of the overall model. The weights of the learning prediction results are implemented by logistic regression and weighted average, and the relationship between the learning model results is implemented by tree model and neural network.

4. A substation equipment failure risk assessment system based on intelligent perception according to claim 3, characterized in that: The base model is verified according to different data subsets, 80% of the data subsets are divided according to the data type and data relevance, and 20% of the data subsets are randomly divided; for the data subsets of the stream data type, the long short-term memory network is used as the base model, and for other types of data subsets, the multilayer perceptron is used as the base model.

5. The intelligent sensing-based substation equipment failure risk assessment system according to claim 3 is characterized in that: The prediction results of the base models will be combined into a new feature matrix. For M base models h1,h2...h m ,According to the prediction results of each base model, a new feature matrix is ​​constructed as the training data of the meta-model.

6. The intelligent sensing-based substation equipment failure risk assessment system according to claim 1 is characterized in that: The fault warning module models fault information, including circuit breaker fault, disconnector fault, transformer fault, arrester fault, capacitor bank fault, control and protection equipment fault, fault caused by environmental factors, fault caused by improper operation and maintenance, line fault, communication fault, and generates a labeled training data set D = {(x1, y1), (x2, y2)...., (x n ,y n )}, where x i is the prediction result of the base model, y i is the corresponding fault label.

7. The intelligent sensing-based substation equipment failure risk assessment system according to claim 1 is characterized in that: The fault warning module first uses the base model to predict the data according to the transmission data of the data acquisition module. After obtaining the prediction results of each base model, it uses the meta-model to calculate the similarity between it and each sample in the training set, and finds the k most similar training samples. According to the labels of these k neighbors, the distance weighted voting method is used to predict the fault information of the new sample.

8. The intelligent sensing-based substation equipment failure risk assessment system according to claim 7 is characterized in that: The similarity is calculated as follows: Among them, d is the prediction result of similarity, m is the prediction result of the base model, is the prediction result of the base model of the collected data, δ is the indicator function, x i To collect data.

9. The intelligent sensing-based substation equipment failure risk assessment system according to claim 1 is characterized in that: The distance weighted voting method is calculated as follows: Among them, N k (x) represents the set of k nearest training samples, w i is the weight of the ith neighbor, c is the corresponding fault type, Indicates the predicted category, y i Indicates the category to which the i-th sample in the training set belongs, w i The calculation method is as follows:

10. The intelligent sensing-based substation equipment failure risk assessment system according to claim 1, characterized in that: The decision support module establishes an operation and inspection strategy library, forms a one-to-one special strategy or a one-to-many progressive strategy according to the complexity of equipment defects and failures, and provides specific maintenance suggestions and priority ranking according to the failure prediction results.

Images